| Nowadays,Human-Computer Interface technology has stepped into the new stage of natural interaction.Sensing technologies play a vital role in this field.Among which the flexible force sensing technology have attracted many researchers for its merits of high density,good flexibility and well correspondence,flexible force sensing technology is an effective and efficiently means for data acquisition of contact force between body surface and surroundings,especially on flexible,curved and even irregular surfaces.Aimed at potentially typical application fields,such as elite sport,national health and rehabilitation therapy,this dissertation carried out a series of practice and explorations,which involve with the calibration of sensor,movement recognition and feedback training method of human motion.The National Natural Science Foundation of China and the National Fitness Science Foundation of the General Administration of Sport of China supported these studies.The entire works was divided into the following three parts:1.Researches on signal calibration and processing method of flexible force sensitive sensors.In the pressure measurement of irregular surface,the shear force and curvature of surface have significant influence on the performance of the flexible force sensor,and the calibration and processing method of sensor signals is studied in this paper.We reformed a modification device of deflection instrument for shear calibration.With this device,the study established a sensor output model of shear stress based on the surface.The study found that shear stress on the surface of the sensor can increase the sensor,which was the path to the vertical loading effect was obvious,close to 8%.We designed a device with variable radius of curved surface for calibration the measurement error caused by the radius of curved surface.With this device,the study established a radius of curvature sensor output model.The study found the increase of the radius of curvature decreased of sensor outputs,which affect the accuracy of the smaller radius surface sensor more obvious.When the radius is 20mm,the output deviation of the sensor can reach-16.7%.The validation experiments shown that the calibration method designed in this paper significantly improved the measurement accuracy of the contact force information and plantar contact force information in the moving environment.2.Researches on the methods of human motion information recognition based on flexible force sensing.The accurate measurement of the contact force of the palm handle in the process of motorcycle driving plays an important role in evaluating the technical rationality of motorcycle riders,optimizing riding techniques and preventing accident risk.We used flexible force sensing technology to realize an acquisition device for contact force information between hand palm and motorbike start on curved touching surfaces.Adaptive fuzzy neural network,particle swarm optimization algorithm was deployed to design the hand function action recognition algorithm.The sensor calibration was combined for processing to optimize the detection effect.The final force sensing glove system was accessible to the common hand operation with real-time,procedure and detail information acquisition.The validation test shown that this motorcycle force sensing glove system worked well.The accuracy of actions identification on the brake and the clutch can reach more than 90%.After calibration,the accuracy of throttle and steering can reach more than 85%.It was accepted and used in the performance diagnosis and training guidance for Australian Superbike Racing League elite athletes.The lower limb strength is the key index of evaluation in competitive sports such as weightlifting and trampoline.In past,force platforms were usually used for lower limb strength assessment.The process of installation and calibration was complex and difficult,and hardly to realize portable and fast measurement.We designed a portable pressure plate with flexible force sensing technology for acquired ground reaction force from feet.Based on Morlet wavelet similarity analysis on the characteristic layer of information in time domain and space domain,a method for lower extremity strength accurate evaluation was proposed.The method had the advantages of convenient installation,less field limits,and so on.The significant of the prediction model was less than 0.01.With this model,this method can distinguish athletes at different levels.This method had been applied to China's trampoline youth talent selection and training guidance.3.Researches on feedback training based on human movement recognition with flexible force sensing technology.Preschool children's ability to balance comprehensively reflected the development of their skeleton and nerve system.The development of low cost,special training instrument to the popularity of the school and the family is of great significance to promote preschool children's physical function.This study integrated balance assessment basing on flexible force sensing recognition and active video games basing on optical body recognition into a proposed scheme of a low-cost balance training integration measurement system.We designed a training prescription management method,which used interest motivation model.We also designed a dynamic adjustment method of training program based on quantitative assessment of balance ability.Compared with the traditional method,this system not only can effectively increase the energy consumption to enhance physical activity level(P<0.05),but also can significantly improve the balance ability(P<0.05).It proved to be a low cost,balance ability for home training and training integration measurement system.Functional electrical stimulation therapy is widely used in the rehabilitation of patients with hemiplegia and other walking disorders.It is the focus of research in this field that combines the independent contraction of target muscles with intensive electrical stimulation during exercise.We used flexible force sensing technology to detect plantar pressure information.Before measurement,the influence of shearing stress such as friction force during walking on the measurement results was suppressed by calibration.Due to the accurate recognition of gait,the electrical stimulation assisted passive muscle contraction and active muscle contraction of brain neural control synchronously.Compared to patients training with conventional electrical stimulation,ipsilateral support phase(P<0.05)and the double support phase(P<0.05)of patients with dynamical electrical stimulation obviously improved.Their exercise capacity proved to be improved.Their brain lateralization the index also improved significantly(P<0.01).The brain function improved implied that significantly improve the rehabilitation effect of patients with hemiplegia. |
PDF Full Text Request